In Spring Data Neo4j 6.0.2 (SNAPSHOT), it seems that bidrectional relationships generate unnecessary statements.
Let's say we have:
@Node
public class EntityA {
@Id @GeneratedValue(UUIDStringGenerator.class)
public String id;
@Relationship("HAS_ENTITY_B")
public EntityB entityB;
}
@Node
public class EntityB {
@Id @GeneratedValue(UUIDStringGenerator.class)
public String id;
@Relationship(value = "HAS_ENTITY_B", direction = Relationship.Direction.INCOMING)
public EntityA entityA;
}
The following code
var a = new EntityA();
neo4jTemplate.save(a);
results in the following statement, which is correct
MERGE (n:`EntityA` {id: $__id__}) SET n = $__properties__ RETURN id(n)
But then if we add a relationship between the two entities:
var b = new EntityB();
a.entityB = b;
b.entityA = a;
neo4jTemplate.save(b);
It seems the relationship is created, then deleted, the created again, which doesn't make sense to me
MERGE (n:`EntityB` {id: $__id__}) SET n = $__properties__ RETURN id(n)
MERGE (n:`EntityA` {id: $__id__}) SET n = $__properties__ RETURN id(n)
MATCH (startNode:`EntityB`) WHERE startNode.id = $fromId MATCH (endNode) WHERE id(endNode) = 804 MERGE (startNode)<-[:`HAS_ENTITY_B`]-(endNode)
MATCH (startNode:`EntityA`)-[rel:`HAS_ENTITY_B`]->(:`EntityB`) WHERE startNode.id = $fromId DELETE rel
MERGE (n:`EntityB` {id: $__id__}) SET n = $__properties__ RETURN id(n)
MATCH (startNode:`EntityA`) WHERE startNode.id = $fromId MATCH (endNode) WHERE id(endNode) = 809 MERGE (startNode)-[:`HAS_ENTITY_B`]->(endNode)
Is this a bug, or can sonmeone explain what is going on? Thanks!
